Three former DeepMind researchers who made a name for themselves by building an AI that beat humans at poker are now applying their skills to the financial markets. Their startup, EquiLibre Technologies, is valued at $500 million after raising a large Series A round from Creandum.

The commonality between poker and finance lies in the applicability of reinforcement learning techniques. The founders, who lack a background in finance, are focused on innovation rather than market efficiency, finding joy in building cutting-edge AI systems.

EquiLibre’s algorithms have been trading billions daily across major stock exchanges since 2025, with a perfect record of positive monthly returns. This success has attracted attention from venture capitalists like Creandum, recognizing the potential in financial markets for AI-driven automation and profitability.

The startup is expanding its compute infrastructure to become one of the largest clusters in Central and Eastern Europe. However, it faces competition from giants like Jane Street, which already uses advanced models and GPUs. EquiLibre’s challenge lies in outperforming rivals while maintaining its unique approach to AI trading.